Learning-related fos-like immunoreactivity in the chick brain: time-course and co-localization with GABA and parvalbumin.

Previous work has shown that, after domestic chicks have learned the characteristics of an object (visual imprinting), there is a learning-related increase in the numerical density of Fos-immunopositive neurons in the intermediate and medial part of the hyperstriatum ventrale, a forebrain region that is a site of recognition memory for the imprinted object. The present study describes the time-course of this effect and has used double-labelling immunocytochemistry to identify neuronal types in which the effect occurs. Chicks were trained by exposure for 1 h to an imprinting (training) stimulus and then given a preference test to determine the strength of imprinting (i.e. of learning). Strongly imprinted chicks were killed 2, 2.5, 3, 3.5 or 4 h (12 chicks in each group) after the start of training and a further group of 12 chicks remained untrained. Sections from the chicks' brains were stained for Fos-like immunoreactivity, and the numerical density of Fos-positive nuclei in the intermediate and medial part of the hyperstriatum ventrale was counted. Relative to untrained chicks, there was a 60% increase in the number of Fos-positive nuclei in the intermediate and medial part of the hyperstriatum ventrale 2 h after the start of training (P = 0.02), but not at any other time. Sections from 10 trained chicks, two killed at each of the above times after training, and from two untrained chicks were stained with anti-Fos antibody as before and also with an antibody against GABA. Approximately 95% of the Fos-positive neurons in the intermediate and medial part of the hyperstriatum ventrale were also immunopositive for GABA. In neurons immunopositive for GABA, there were significantly (P = 0.02) more Fos-positive nuclei in the intermediate and medial part of the hyperstriatum ventrale 2 h after the start of training than in untrained chicks. Five chicks killed 2 h after training and five untrained chicks yielded sections for the next experiment; sections were double labelled for (i) Fos and (ii) either Calbindin-D28k or parvalbumin. Training gave rise to a significant (P = 0.017) increase in numerical density of Fos-positive nuclei of neurons that were immunonegative for Calbindin-D28k. This increase occurred in neurons that were immunopositive for parvalbumin. The use of alternative antibodies for GABA, Calbindin-D28k and parvalbumin in trained and untrained chicks confirmed the double-staining pattern observed in the quantitative experiments. The results demonstrate that the learning-related increase in Fos-like immunoreactivity following training is transitory and have localized the increase to a population of neurons immunopositive for GABA and parvalbumin, but not Calbindin-D28k.